The Fuel and Engine Bible - how engines work including 2 stroke, 4 stroke and wankel (rotary) engines, fuel, octane rating, power, bhp, gas types and grades, carburettors, fuel injection, tuning, tweaking, nitrous, turbos, superchargers, chipping, hybrids, how to keep your engine running at peak fitness and much more.

Gas-mileage, mpg and why American cars can never match the EPA estimates

Gas-mileage is the quickest indicator of how efficient a car is in terms of fuel used for distance driven. Engine size and power, driving conditions, weather (wind especially) and vehicle weight all affect mpg. Measuring gas-mileage is really easy but it's surprising how many people don't know how to do it. Basically, zero your trip counter next time you fill up, then drive as normal. When you fill up again, let the petrol pump fill to the auto-cutoff point and then make a note of the trip meter reading. Gas mileage is the number of miles on your trip meter divided by the number of gallons the petrol pump put into your tank. You'd be surprised the number of people who use the manufacturer figure for the size of the tank in that calculation instead of the amount of petrol actually put in.
In England and Europe, pumps deliver in litres, so in the UK it's miles-per-litre, although most advertising still uses miles per gallon. It's worth noting that an English gallon is 1.2 US gallons. So when you see a car in England that advertises 40mpg, it's the equivalent of 33mpg in the US.
In the rest of Europe it's normally advertised as litres per 100 km. So for example, 28mpg (UK) is about 10litres/100km. Often this is short-handed to 1-in-10, meaning 1 litre used in 10km of driving.

The EPA

The American EPA (Environmental Protection Agency) rates all cars sold in America with gas-mileage figures, advertised as EPA-rated mpg figures on the new car sticker. It's one of the things car manufacturers rely on to sell their vehicle, especially with today's high fuel prices. Not many people understand this, so I'm here to take some of that confusion away and tell you what the EPA figures really mean.

First of all, there's the sticker you'll see in every new car in an American showroom, an example of which is seen on the right. There's a load of technical blurb on there to advertise the vehicle, but the two big numbers are the EPA-certified fuel information figures. In this case 20mpg city and 28mpg highway. So you see these figures and you get into your head a rough idea of how often you'll be filling up. The problem is that these are very rough estimates. If you read the small print, it says this:
"Actual mileage will vary with options, driving conditions, driving habits and vehicle condition. Results reported to EPA indicate that the majority of vehicle with these estimates will achieve between 17 and 23mpg in the city and between 23 and 33mpg on the highway."
Okay so it's pretty obvious that driving habits, conditions and vehicle options (like a bloody big roof rack) will affect your mpg, but what's less obvious is the "between" figures. It's basically a get-out clause. In this example, the vehicle is more likely to get 17mpg in the city and 23mpg on the motorway - the low end of the "between" figures.
In the 1980s, the EPA conducted a study on their results vs. the real world, and discovered most drivers got significantly lower mpg figures than the EPA predicted. As a result, EPA estimates on the new car labels were dropped by 10% for city and 22% for highway from their actual results. In 2006 they dropped another 8% from those figures again to try to make the numbers match more closely.
Even that isn't the end of the story though. What you really need to know is how the EPA come up with the figures in the first place. Before you carry on, you might want to put down any drinks or breakables because I know what your reaction will be at the end of this. Ready?
Congress and car company lobbyists require the EPA to measure mpg figures using the following simulated real world conditions in a lab. That's right - EPA testing happens on a dyno in a lab, not on the open road.

• Average highway speed : 49mph
• Maximum highway speed : 60mph
• Temperature : 75°F
• No rapid acceleration
• No air conditioning
• No passengers
• No rough roads
• No hills
• No wind
• No low tyre pressures
• No ethanol in gas

Well the first problem is the last point : no ethanol in gas. In America, you can't buy zero-ethanol petrol - it's all E-10 (see above) so you're already going to be down 5% on the EPA figures even if you could meet all the other requirements. And for the love of God, who drives like this? 49mph on the motorway? Maximum speed 60mph? Perhaps when the model-T Ford was the Big Thing, these were valid speeds, but nowadays (and by 'nowadays' I mean 'in the last 6 decades') motorway speeds are typically 70mph maxing out at 90mph (if you're in Europe anyway). What about the rest of it - no hills, no passengers, no rough roads? Have the EPA actually driven a vehicle in the real world recently?
As a rough benchmark, driving at 65mph instead of 49mph will decrease mpg by 20%. Driving at 75mph will take another 25% off that. In short, you should pay very little attention to the EPA estimates because they are, for the most part, completely meaningless.
Muddying the waters even further
Remember above I said that the city and highway figures were "between" figures, or the average of the high and low EPA tests? Well I'll give you one guess which figure the car manufacturers use in their print and media advertising. That's right - the high-end of the range. In the example above, the low highway figure was 23mpg and the high was 33mpg. In this case, the advertising will always publicise the 33mpg figure. You will likely get not much more than the low city figure - 17mpg.
Trying to give you a concise answer.
They say a picture speaks a thousand words. I don't have a picture for you but I do have a table. This is a quick reference for you to show all the various figures that go into the EPA estimates, the advertising and what you should expect in the real world. It's based on the Mercedes CLK320 sticker shown above. The blue row shows what you'll see on the EPA sticker in the window of the car. The red shows the figure you'll see on TV and the green row shows what you should expect when you drive this car in the real world.

	City Low	City High	City Avg.	Highway Low	Highway High	Highway Avg.	Combined Avg.
EPA LAB TEST	21.6	26.3	23.9	26.3	42.1	34.2	29
-15% (1980 correction)	18.4	25	21.7	25	35.8	30.4	26
-8% (2006 correction)	17	23	20	23	33	28	25.5
-5% (you're using E-10 petrol)	16.1	21.8	19	21.8	31.3	26.6	25.1
What you should expect	15	20	17.5	20	30	25	21.2

If you're curious about what others are getting in the real world, there's three websites that will help you out:
fueleconomy.gov is the US government's own website where people like you and I contribute to their real-world mileage database.
GreenHybrid.com which is based more on hybrid vehicles, and TrackYourGasMileage.com. The following graphic is continuously updated based on my account at trackyourgasmileage.com. Compare my actual mileage to the claimed 21/24 that the EPA advertises for my vehicle. If you're interested in nauseating detail, clicking the graphic will take you to a tank-by-tank breakdown:

So the EPA numbers are essentially useless then?

Yes, apart from for one thing. Too many people try to perpetuate the myth that the EPA values are intended to suggest what a driver could expect to get in the real world. As I've shown in mind-numbing detail above, this is simply not the case. Instead, they are best used as a comparison between one vehicle and another, ie if one vehicle is EPA-rated at 20mpg and another is EPA-rated at 25mpg, then you can pretty safely conclude that the latter gets 25% better mileage than the former, and nothing else. For a good read on this subject see the Patrick Bedard column in the Feb 2006 issue of Car and Driver magazine.

2008 : The EPA adjust their figures

So you've waded through all this crap about the EPA and how the estimates are pretty much irrelevant, only to come across the above heading. Yes, it's true. The EPA changed the way they measure mpg figures starting in 2008. The big changes are:

• The new tests start with the car at 20°F. The old tests started at 75°F. Why the change? A cold vehicle uses more energy than a warm one. Cold temperatures are especially hard on the batteries that power hybrids, so hybrid mpg ratings dropped.
• The new tests use a maximum highway speed of 80mph instead of 60mph. At that speed more engine work is done to overcome drag than to actually keep the vehicle moving at speed.
• The new test includes hard acceleration. The old test used gentle acceleration. This one also affects hybrids because hard acceleration relies entirely on the regular petrol engine and uses none of the electric hybrid parts.
• The new test now assumes air conditioning is used 13% of the time. The old test didn't use air conditioning at all. 13% is the mean average for all major cities across all times of year for the US.

But even with these new standards, the EPA test still takes no account of hills or wind. This has the effect of skewing the test in favour of larger vehicles like SUVs. If hills and wind were included, the results would be radically different - larger, heavier vehicles use more energy to travel into wind and up hills, ie. more fuel. The 2008 EPA estimates would be far more useful if they included these factors. Because they don't the overall fuel consumption figures for SUVs are lower than is realistic.
For example. Assume a car gets 40mpg without the hills and wind test, and 38mpg with. Now imagine an SUV doing the same tests gets 24mpg without the hills and wind, and 19mpg with. For the sake of comparison the car's 40mpg vs. the SUV's 24mpg doesn't look as bad as the car's 38mpg vs. the SUV's 19mpg. But I digress.

So what's the outcome of this? Well first of all, the new figures are closer to what you or I could get from a car. Hybrids still suffer a hit of about 30% loss of mpg for highway and 20% for city. Regular non-hybrids will drop about 12% for highway and 8% for city. It's a lot closer to the real world than it was, but is it close enough or should we still just use the EPA figures as an arbitrary comparison of vehicle mpg as measured against an arbitrary scale? Time will tell, but it could be argued that the major car manufacturers and oil companies lobbied for this change to take the shine off hybrid vehicles - after all, they're the ones that suffer the most with the new rating. If you're curious, here's an example of the 2008 window sticker.

The transatlantic conundrum

Here's a question for you : why do identical cars, made by the same manufacturer, get less mpg in America than they do in Europe? I know a lot of you are reading this now thinking "Aha - that's because an imperial gallon is larger than a US gallon". Yes, but even adjusted for that, it's still true. Take for example the 2008 Honda Civic 1.8 i-VTEC 5-door manual. It's a good example of an average family saloon/sedan car. The trim levels are identical, as are the engine and gearbox, and power and torque figures. Oddly, the European cars weigh more for the same trim level. The following are all converted to US gallons:

	City mpg	Highway mpg	Combined mpg	Peak power	Peak torque	Kerb weight
UK	28	43	36	140hp @ 6300rpm	128lb.ft @ 4300rpm	1281kg
US	26	34	29	140hp @ 6300rpm	128lb.ft @ 4300rpm	1241kg

Another example saloon / sedan car. The 2008 Volvo S40 2.4i. Again - same trim level, engine and gearbox. Once again, the European car weighs more:

	City mpg	Highway mpg	Combined mpg	Peak power	Peak torque	Kerb weight
UK	20	37	28	170hp @ 6000rpm	170lb.ft @ 4400rpm	1481kg
US	20	28	24	168hp @ 6000rpm	170lb.ft @ 4400rpm	1460kg

Doing the test differently

Interestingly, the EU mpg test is done on the road instead of as a lab test so you would expect it to give a worse figure, not a better one because the test isn't skewed in favour of unrealistic scenarios. However, it's obvious when you look at the 2008 EPA testing methodology that they took a leaf out of the EU book. This is how the EU do their mpg tests:

• Urban (city) : cold engine, accelerations, steady speeds and decelerations. The average speed is 12mph representative of city commute speeds (very harsh on mpg) and the distance is 2.5 miles.
• Extra-Urban (highway) : warm engine, accelerations, steady speeds (50% of the test) and decelerations. Max speed is 75mph, average speed is 39mph, distance is 4.3 miles.

So why the huge difference? The EU regulatory body and the EPA use very similar methods of determining the mpg figure so it's not that. You'd expect the figures to be within a couple of percent of each other, but they're clearly not. In fact with the EU cars weighing more you'd expect their figures to be worse. If you think you know the true answer to this (ie, not Big Oil conspiracy), drop me a line. Suggestions so far:

Different Petrol Octane/Composition

Good idea but the calorific value of low-octane fuel is pretty much the same as high-octane.

European and Japanese cars are designed for higher-octane fuels (higher compression).

Might be true but most European vehicles run Euro95 petrol, which is the equivalent of American (R+M)/2 89 octane. Plus, the engine specs are identical - same compression ratio, same torque, same horsepower.

Different Oil Type/Viscosity used

Different 'Map' on the ECU for different emissions laws

Tyres - different rolling resistance?

For my money, the best one is the different engine map in the ECU for emission laws, although emission laws are stricter in a lot of parts of Europe than they are in the US which you'd think would make the mpg figures worse.

Reader ideas

The Transatlantic Conundrum has generated more buzz on my email than Lindsay Lohan getting out of a limo with no knickers on. Some of the ideas are quite sensible. Some are way out in left field. Scott Brereton emailed me with one of the more intelligible ideas:
Like yourself I suspect that its related to engine maps, and I think it may be a quite subtle effect of differing emissions laws.
After doing some research it turns out that in the UK emissions standards measure carbon monoxide(CO), hydrocarbons(HC) and lambda. Looking at the regulations for the US, as far as I can tell CO, HC and nitrogen oxides(NOx) are measured. Due to differing test methodologies, I can't make a direct comparison between the CO and HC figures, hence why I'm phrasing all this with lots of 'maybe's.
Since NOx is formed under lean combustion conditions, it might be the case that fuel maps in the US are tuned richer than in the UK to minimise NOx production, this will lead to higher CO and HC but if we suppose that the CO and HC standards required in the US are more relaxed then those in the UK this would not be a problem. This richer fuel map might go some way to explaining the differing fuel consumption figures seen on either side of the Atlantic.

Reader Blaine writes: It's based in the US EPA's pre-occupation with NOx levels. Recall that European smog police pay little attention to NOx, and concentrate on HC, CO, and CO2. Also recall that NOx formation occurs at extremely high temps (2700+ degrees?).
US emissions systems use a catalyst to bring down NOx levels. The problem with this catalyst is that it doesn't work at "low" temperatures - even those typically found in exhaust systems. US emissions systems make this catalyst work by dumping raw fuel (or running an exceedingly rich mixture) into the exhaust stream, to burn in this NOx catalyst to keep it hot enough to perform the reduction reaction to eliminate/reduce NOx levels. This results in somewhat excessive fuel consumption, obviously.
Tuners in the US, on OBD-II vehicles, have figured out that eliminating this fuel enrichment can result in a fairly substantial gain in fuel economy, especially when combined with fuel-map and timing tweaks designed to increase fuel economy in other circumstances.
It's the "good-fast-cheap" triangle: "good, fast, cheap; pick two". Automotive engines are governed by a similar triangle: power, economy, emissions. European ECM's can pick two; American ECM's can only pick one, by federal mandate - emissions (every other consideration - power, fuel economy, driveability, etc - comes in as a distant second place). By going to a more european style of ECM mapping, American tuners can work some pretty amazing feats, and still maintaining emissions levels.

Torque and BHP

There are two values that are generally published for an engine which tell you how strong the engine is, and they are torque and bhp - brake horsepower.
Torque is a measure of the twisting power of the engine. Torque is directly related to acceleration; the more torque, the quicker you'll get up to speed. Horsepower is what will keep you at speed once you've accelerated and is directly related to the torque readings. So a high-torque, low-horsepower engine will accelerate well but be unable to maintain a high speed. Similarly, a low-torque, high-horsepower engine will not have much acceleration but will be able to go at a fair clout once it's going.

The difference between horsepower and bhp

In England and the US, horsepower means Imperial horsepower. The technical definition of this is "the power a horse exerts in moving 550 pounds of cargo a distance of one foot in one second." This calculation can include just the horse and its own weight. Horsepower can be defined many ways. One horsepower equals 746 watts, and as such, proper SI units are normally used instead. The term horsepower is more a legacy term than anything else.

The term brake horsepower came about because of an apparatus called a water brake that can be used to measure horsepower. Today all manner of brakes are used from hydraulic to electrical. They all perform the same function though, and that is to load up the engine and measure the torque with strain gauges. BHP figures can be calculated from the measured torque values to determine the power of the engine at any given rotation. If bhp figures are published without any other data, you've got to assume they're measured at the crank. The problem is that once you add on clutches, flywheels, gears, driveshafts and all the other components between the engine and the wheels, the actual power at the wheel is often noticably less. So sometimes you'll see bhp figures noted as "at the wheel". This means the torque has been measured with the wheel being turned through all the above connections to give a more accurate power reading.
In the bad old days, bhp readings would be taken with the engine running in "optimum" condition, ie. with no oil or water pumps attached, direct cold-air injection, super-cooled coolant, no exhaust back pressure or catalytic converters and so on and so forth. Fortunately today there are standards that have to be maintained. Most recently, in 2005 the SAE made some changes to the test procedures to eliminate some of the 'slop' in power measurements, and for car manufacturers to be able to make valid SAE-certified bhp claims, their tests must now be monitored by SAE representatives. The results of this change were interesting if only because the bhp values for engines changed without the engines themselves being modified. For example the Honda Element engine remained exactly the same, but its bhp rating dropped from 170bhp to 165bhp, simply because of the new procedures.
It's worth pointing out that whilst the rest of the world used bhp or kW (kilowatts) to publish power figures for engines, in America they typically used to use hp(SAE) instead, meaning the rated power of the engine as installed in the vehicle, ie including all the engine components, pumps, drivetrain etc. Having said that, even today, all hp(SAE) or SAE-certified bhp figures are taken at the flywheel and thus still don't really tell you how much power is getting to the wheels. The only way to know that is to put your car on a dynomometer (a dyno) and get true at-the-wheel readings.

Calculating horsepower and bhp

The formula to calculate bhp from a given torque reading is as follows:

Pi is obviously 3.14159, the torque value should be in pounds-feet and RPS is revolutions per second - RPM/60. So do a little elementary maths and you can massively simplify the formula down to this:

The formula to calculate regular horsepower from a given torque reading is as follows:

Pi is still 3.14159, but this time the torque value should be in newton-metres. Again, simplified the formula becomes:

Horsepower and altitude

A little footnote here : assuming all other factors are the same - same octane of petrol, same filters, same weight of vehicle etc., you lose about 3% of your engine's horsepower for every 300m / 1000ft above sea level that you drive. Worth knowing next time you go carving up the Stelvio Pass in Italy. At 9000ft, you're a full 27% down on engine power compared to sea level.